“…In the confirmatory phase we tested three models: (a) a unidimensional model that tests the hypothesis that loneliness is a heterogeneous construct composed of different types of subjective experiences of social interaction (de Jong Gierveld & Kamphuis, 1985;Moorer & Suurmeijer, 1993;van Tilburg, 1988); (b) an oblique model that considers that social and emotional loneliness are related theoretically and empirically (Dykstra & Fokkema, 2007;Leung et al, 2008, van Baarsen et al, 1999van Baarsen et al, 2001), and (c) an orthogonal model that considers the existence of two unrelated types of loneliness (Kunts et al, 2010;Zammuner, 2008). To determine which model best fitted the data, we examined the goodness of fit indices available in AMOS for tetrachoric correlations: RMR .08 (Hu & Bentler, 1998), GFI and AGFI .90 (Hoyle & Panter, 1995), NFI 0.95 (Bentler, 1990) and PNFI .90 (James, Mulaik, & Brett, 1982).…”